Bus Backfeed options - 120% rule

[peakbagger]

I am getting ready to install an addtional 2 KW grid tie system and want to avoid a line side tap. The local area really doesnt have any master electricans with solar experience so I need to work with one and point him in the right direction . The local inspector has no electrical background and basically has told me that a master has to sign off on the install. The area uses NEC 2011

Here is what I have

200 Amp breaker panel, there is no bus rating so I will assume it is 200 AMP bus

One 2 KW grid inverter - 8.35 max amps hooked to a 2 pole 15 amp breaker

One 1 KW grid inverter - 8.35 max amps hooked to a 1 pole 15 amp breaker

New 2 KW grid tie inverter - 8.35 max amps hooked to 2 pole 15 amp breaker.

The breaker amperage ratings are all the manufacturers recomended sizing and down sizing to 10 Amp breakers would exceed the 80% continuous rating of the breaker

120% backfeed is 40 AMPs maximum based on breaker ratings, If I look at the potential load per bus leg , I have an unbalanced backfeed with 30 Amps on one leg and 45 amps on the other. I assume this will not fly?

I do see references to 705.12(D)(2) Exception for battery sourced multimode inverters, that is real tempting as then all I have to do is add up the inverter ratings and multiply by 1.25. This yields 31.3 and 20.8 amps. Looks like this works but since I have battery less units I expect this wont fly ? Arguably either unit is limited on its output by internal circuitry but the code is the code


Options
All three arrays have different exposures, I wouldnt mind buying the new inverter with dual MPPT but they seem to have fallen out of favor.

Install a 110 to 220 volt step up transformer on the 1 KW unit and switch to a smaller 2 pole breaker? This balances the load and gets me below 40 amps if I install a 10 amp 2 pole breaker. At that point I would probably do a line size tap due to the cost and loss of the transformer.

Install a subpanel for PV only, have it feed a 40 amp breaker in the main panel.? This wouldnt be so bad but I am not sure the code allows this?

So anyone have any suggestions or thoughts besides shutting down the 1 KW system?

[inetdog]

Install a subpanel for PV only, have it feed a 40 amp breaker in the main panel.? This wouldnt be so bad but I am not sure the code allows this?

So anyone have any suggestions or thoughts besides shutting down the 1 KW system?

As long as the subpanel meets the 120% rule also (1.2 x bus rating greater than 45 + 40 main if you can find one.), that is fine. A 100A bus subpanel with a 50A or 75A main and three 15A feeds would be OK since the total is less than 120A.
And the main panel is OK with all of the back feed coming through the single 40A breaker.
Unfortunately the code does not let you just add the maximum power output of the GTIs when looking at the 120% rule.

Oddly enough, if the inverters were hybrid GI instead of just GTI, you would be able to add up 125% of their individual rated GI outputs (1.25 x 8.35 = 10.44 each) instead of adding up the breaker sizes, and you would not have a problem.

[bcroe]

Exposure

You say your arrays have different exposures? Does that mean, different orientation,
with different times of peak power? If 2 arrays were the same voltage and peaked at
widely different times, you could parallel them to one inverter, which would have less
capacity than the 2 separately. That would reduce the apparent total to closer to the
actual possible power. Bruce Roe

[peakbagger]

I have always assumed that MPPT works the best when the panels are identical with the same exposure. I am not sure I want to risk impacting the inverter efficiency bty having two strings with widely different exposures. Plus one string is on a pole mount vs the new string is on a roof so the temps will be quite different.

[bcroe]

120%

I have always assumed that MPPT works the best when the panels
are identical with the same exposure. I am not sure I want to risk impacting the inverter
efficiency bty having two strings with widely different exposures. Plus one string is on a pole
mount vs the new string is on a roof so the temps will be quite different.

I'm not convinced your step up transformer idea would work. It would raise the impedance the
inverter sees at the line, possibly tripping the anti island function. It would be more likely to work
(in this regard) and be twice as efficient as a 120V- 240 autotransformer.

Your secondary panel idea allowing a single breaker to your main, sounds attractive if it doesn't
break any rules. Certainly less expensive than buying an inverter.

My paralleled arrays idea fails if there is significant overlap of peak output times.

My rule of thumb, is the MPP of a properly operating array is about 78% of Voc. This voltage
doesn't vary a lot over the range of useful power output, so strings with the same Voc but
different sun intensity can still be efficient in parallel. If the instantaneous Voc were to shift
on one string due to temp, a common MPP output would be reduced a somewhat less percentage.

If a panel in a string were shaded, bypass diodes could allow the string to still produce useful
power at a lower voltage. But if another string in parallel continued with no shade, a common
MPP inverter would have to choose between running at much reduced voltage for all, or running
at full voltage for the unshaded string and receiving very little current from the string with a
shaded panel. Either way would reduce power much more than having separate MPP inverters
for each string. The argument can be extended to a micro for each panel. Bruce Roe

[Naptown]

What you are describing on the sub panel works. the Sub panel would not be limited to the 120% rule. This panel should be clearly marked and is actually more of a combiner box than anything

[inetdog]

I have always assumed that MPPT works the best when the panels are identical with the same exposure. I am not sure I want to risk impacting the inverter efficiency bty having two strings with widely different exposures. Plus one string is on a pole mount vs the new string is on a roof so the temps will be quite different.

If the two strings are within 10% with respect to Vmp during the time that they are both producing significant power, the output of the inverter should be within about 5% of the sum of the two strings operated at their individual Vmp. But this general guideline affects primarily use of different types of panels.
The only effect which can cause changes in the Vmp with identical panels will be temperature and partial shading. The temperature effect will not be enough to cause even a 5% difference in Vmp.

The exposure of the panels will change Imp but will have a negligible effect on the Vmp at which that current is obtained. Xantrex did some experimental studies that confirmed that paralleling arrays that differer only in exposure will not cause any loss of power, and that any clipping loss caused by an overlap of the two arrays exceeding the capacity of the inverter for a short time at midday will be more than made up for by the extended hours of higher power operation.

[bcroe]

120%

What you are describing on the sub panel works. the Sub panel would not be limited to the 120% rule. This panel should be clearly marked and is actually more of a combiner box than anything

I think you are saying, it doesn't apply to the sub panel. But if the sub panel connects to a breaker on
the main panel, is that breaker still limited to 20% of the main? Bruce Roe

[inetdog]

I think you are saying, it doesn't apply to the sub panel. But if the sub panel connects to a breaker on
the main panel, is that breaker still limited to 20% of the main? Bruce Roe

There is a good argument as to the justifiability of applying the 120% rule to a subpanel which is only serving as a combiner, but particularly if there are other loads in the subpanel, it is technically governed by the same bus loading rule that applies to a main panel. It is just usually easier to meet that rule.
The saving feature that this sort of subpanel combiner has is that the bus calculation at the main panel can be based not on the individual subpanel breakers or even the breaker at the main panel, but rather the main breaker at the subpanel.
So you could take three 8A GTIs, run each of them through a 20A breaker into a subpanel with a 30A main breaker and connect that subpanel feeder to a 50A breaker in the main. The calculation at the main would then use the 30A figure and not 60A or 50A.

Also, be careful with your words. A single backfed breaker at the main, without a subpanel is not limited to 20% of the main breaker size.
The sum of the backed breaker(s) and the main must not exceed 120% of the bus rating. Even if you buy a panel with a pre-installed main breaker, you cannot assume that the bus rating of that panel is equal to the main breaker size.
Manufacturers often use a 150A bus in their 100, 125 and 150A panels. And in some cases the bus rating of a 200A panel will be 225A.

If you have a 200A main, with a 200A bus, you can still replace the 200A main with a 150A main and then be able to use up to 90A of back feed (240A - 150A).

[bcroe]

Bus size

Also, be careful with your words. A single backfed breaker at the main, without a subpanel is not
limited to 20% of the main breaker size.
The sum of the backed breaker(s) and the main must not exceed 120% of the bus rating. Even if
you buy a panel with a pre-installed main breaker, you cannot assume that the bus rating of that
panel is equal to the main breaker size.
Manufacturers often use a 150A bus in their 100, 125 and 150A panels. And in some cases the bus
rating of a 200A panel will be 225A.

If you have a 200A main, with a 200A bus, you can still replace the 200A main with a 150A main
and then be able to use up to 90A of back feed (240A - 150A).

Looking at my own Sq D QO main panel, each bus is 1" wide by 1/4" thick, pretty substantial. I
wonder how their current rating can be found?

The whole 120% idea is of course idiot proof. But the only way to reach current of the main +
backfeed breaker in the bus, is put the 2 at one end, then put loads equaling more than 120%
along the other end. That isn't sustainable, because when the backfeed stops, the main will drop.
My loads certainly don't come anywhere near the capacity. A cure is to put the backfeed at the
opposite end of the bus from the main. Then there is no way the source currents can be additive,
feeding loads in the middle. Bruce Roe

[inetdog]

Looking at my own Sq D QO main panel, each bus is 1" wide by 1/4" thick, pretty substantial. I
wonder how their current rating can be found?

The whole 120% idea is of course idiot proof. But the only way to reach current of the main +
backfeed breaker in the bus, is put the 2 at one end, then put loads equaling more than 120%
along the other end. That isn't sustainable, because when the backfeed stops, the main will drop.
My loads certainly don't come anywhere near the capacity. A cure is to put the backfeed at the
opposite end of the bus from the main. Then there is no way the source currents can be additive,
feeding loads in the middle. Bruce Roe

You cannot tell by looking, but if there is a part number you can check with the manufacturer. Or you can just check the part number of the whole panel with the manufacturer to see what ampacity bus is used in it.

And the code already requires you to put the backfeed breakers at the other end of the bus from the main. (That is why they give you the extra 20%!)
It does not make complete sense, since the bus overload will, as you say, occur only at the exact point along the bus that the 120% load connects. But that is the way it is written.

There is no guarantee that the back feed will stop when you apply a 120% load (not a short circuit), so you can overload that one spot on the bus for as long as you want as long as the load does not exceed the sum of the two feeds.
In theory, if the bus rating is based in part on the ability of the bus to transfer heat to unloaded parts of the bus, the overloaded contact point (with contact resistance) between the load breaker and the bus could overheat. (I am just playing devil's advocate here, not saying I agree with the rule.)

[peakbagger]

Its a 25 year old Westinghouse main panel so I dont hold up much hope on getting any specs on the bus bar rating on the web . Down rating the main breaker is intriguing. The biggest load in the house is an oven and on rare occasions a welder but I expect I could drop down to a 150 AMP main breaker and never notice it. Given the current real estate around the existing panel, fitting in an auxiliary panel would be difficult or might require moving the main panel. If I had to move the main panel, I would probably just buy a new one with 250 AMP bus and a 200 Amp main breaker. Of course since I have to work with an electrician for a sign off, it may all come down to what he is comfortable with, nevertheless armed with options I can challenge him if he goes of on an expensive tangent.

Thanks for all the input!

[Naptown]

I think you are saying, it doesn't apply to the sub panel. But if the sub panel connects to a breaker on
the main panel, is that breaker still limited to 20% of the main? Bruce Roe

Correct the total of all. But the governing ampacity is the maximum output from the inverters not necessarily the total of the breaker ampacity.
Remember that inverters and solar are current limited so they will never produce more than the Isc amperage as opposed to a POCO line that could go up into the thousands of amps.

[inetdog]

Correct the total of all. But the governing ampacity is the maximum output from the inverters not necessarily the total of the breaker ampacity.
Remember that inverters and solar are current limited so they will never produce more than the Isc amperage as opposed to a POCO line that could go up into the thousands of amps.

I thought so too until I read the exact wording of the code (2011):

(2) Bus or Conductor Rating. The sum of the ampere ratings of overcurrent devices in circuits supplying power to a busbar or conductor shall not exceed 120 percent of the rating of the busbar or conductor.

Exception: Where the photovoltaic system has an energy
storage device to allow stand-alone operation of loads, the
value used in the calculation of bus or conductor loading
shall be 125 percent of the rated utility-interactive current
from the inverter instead of the rating of the overcurrent
device between the inverter and the bus or conductor.

Nothing about using the rated maximum current for a non-battery GTI.

Remember that the power input from the panels is limited, but depending what kind of output voltage range the GTI has, the current may be more than that corresponding to the input power at the nominal output voltage.

[Naptown]

I thought so too until I read the exact wording of the code (2011):

(2) Bus or Conductor Rating. The sum of the ampere ratings of overcurrent devices in circuits supplying power to a busbar or conductor shall not exceed 120 percent of the rating of the busbar or conductor.

Exception: Where the photovoltaic system has an energy
storage device to allow stand-alone operation of loads, the
value used in the calculation of bus or conductor loading
shall be 125 percent of the rated utility-interactive current
from the inverter instead of the rating of the overcurrent
device between the inverter and the bus or conductor.

Nothing about using the rated maximum current for a non-battery GTI.

Remember that the power input from the panels is limited, but depending what kind of output voltage range the GTI has, the current may be more than that corresponding to the input power at the nominal output voltage.

Even so with a sub panel arrangement it is possible to limit the OCPD to lower than the 120% rule. The issue is his 1 kw system puts a little over 4A but has a 15A circuit breaker.
In this case a 6 or 8 circuit solar combiner/sub panel to combine all the arrays will allow him to connect all of them to 1 2 pole 40A breaker and feed the 200A panel no problem.